How Much Reconstruction Does Quantum Machine Learning Need? Late Fusion of Independently Trained Quantum Subcircuits
作者: Prabhjot Singh, Adel N. Toosi, Rajkumar Buyya
分类: quant-ph, cs.DC, cs.LG
发布日期: 2026-08-06
备注: 11 pages, 6 figures. Includes technical appendix
💡 一句话要点
提出晚融合方法以降低量子机器学习的重构成本
🎯 匹配领域: 支柱九:具身大模型 (Embodied Foundation Models)
关键词: 量子机器学习 电路切割 晚融合 多模态学习 鲁棒性
📋 核心要点
- 现有的量子机器学习方法在电路切割后重构输出时面临指数级的采样开销,导致运行效率低下。
- 本文提出的晚融合方法通过独立训练子电路并用经典方法组合输出,显著降低了重构成本。
- 实验结果显示,晚融合方法在准确性上与完全重构相近,但成本更低且对噪声更鲁棒。
📝 摘要(中文)
电路切割允许大型量子神经网络(QNN)在小型设备上作为独立子电路运行,但重建输出的过程在切割数量上会带来指数级的经典采样开销,成为先前工作的主要运行成本。本文提出了一种晚融合方法:每个子电路独立训练和测量,使用一个小型经典头部组合其输出,这种决策级的线性组合借鉴了多模态学习。通过引入量子度量拨盘Q和切割纠缠诊断,本文表征了重构需求的权衡。实验结果表明,独立训练的晚融合方法在每个控制点上与完全重构的准确性相差仅0.04,并且在成本上显著降低,同时对噪声更具鲁棒性。
🔬 方法详解
问题定义:本文旨在解决量子神经网络在电路切割后重构输出所需的高昂经典采样开销,现有方法在处理此问题时效率低下。
核心思路:提出晚融合方法,允许每个子电路独立训练和测量,使用小型经典头部进行输出组合,从而避免重构过程中的高成本。
技术框架:整体架构包括独立训练的量子子电路、经典融合模块和量子度量拨盘Q。每个子电路在小型设备上独立运行,最终通过经典方法进行输出融合。
关键创新:引入量子度量拨盘Q和切割纠缠诊断,能够量化不同任务对重构的需求,提供了一种自我表征的替代方案。
关键设计:在实验中,控制了重构预算Q的设置,采用Spearman相关系数评估重构需求,并在多种数据集上进行了验证。
🖼️ 关键图片
📊 实验亮点
实验结果表明,独立训练的晚融合方法在每个控制点上与完全重构的准确性相差仅0.04,同时在成本上显著降低,且对设备噪声表现出更强的鲁棒性。这一方法在104次实验中达到了Spearman相关系数ρ=0.59,验证了其有效性。
🎯 应用场景
该研究的潜在应用领域包括量子机器学习、量子计算和多模态学习等。通过降低量子神经网络的重构成本,晚融合方法可以在资源受限的环境中实现高效的量子学习,推动量子技术在实际应用中的发展。
📄 摘要(原文)
Circuit cutting lets a large quantum neural network (QNN) run as independent subcircuits on small devices, but rebuilding its outputs by reconstruction carries a classical sampling overhead exponential in the number of cuts - the dominant runtime cost in prior work. We ask whether, for machine-learning tasks, this step is necessary, and replace it with late fusion: each subcircuit is trained and measured independently, and a small classical head combines their outputs - a linear-cost, decision-level combination borrowed from multimodal learning. To characterize the trade-off we introduce a quantumness dial $Q$, a tunable reconstruction budget interpolating from pure fusion to full reconstruction, and a cut-entanglement diagnostic that indicates how much reconstruction a task needs (Spearman $ρ=0.59$ over $104$ runs). Across synthetic and standard datasets, independently trained late fusion matches full reconstruction accuracy within $0.04$ at every point of the controlled sweep and on every classical benchmark, at exponentially lower cost; it is also markedly more robust to shot and device noise. Controlled entangled-data experiments locate the boundary where fusion must fail. We do not claim advantage over classical machine learning - consistent with recent benchmarking, quantum offers no accuracy edge on these datasets. Late fusion is thus an efficient, noise-robust, self-characterizing alternative to reconstruction for circuit-cutting QML.